Molecular epidemiology has been central to uncovering P. knowlesi as an important cause of human malaria in Southeast Asia, and to understanding the complex nature of this zoonosis. Species-specific parasite detection and characterization of sequences were vital to show that P. knowlesi was distinct from the human parasite species that had been presumed to cause all malaria. With established sensitive and specific molecular detection tools, surveys subsequently indicated the distribution of P. knowlesi infections in humans, wild primate reservoir host species, and mosquito vector species. The importance of studying P. knowlesi genetic polymorphism was indicated initially by analysing a few nuclear gene loci as well as the mitochondrial genome, and subsequently by multi-locus microsatellite analyses and whole-genome sequencing. Different human infections generally have unrelated P. knowlesi genotypes, acquired from the diverse local parasite reservoirs in macaques. However, individual human infections are usually less genetically complex than those of wild macaques which experience more frequent superinfection with different P. knowlesi genotypes. Multi-locus analyses have revealed deep population subdivisions within P. knowlesi, which are structured both geographically and in relation to different macaque reservoir host species. Simplified genotypic discrimination assays now enable efficient large-scale surveillance of the sympatric P. knowlesi subpopulations within Malaysian Borneo. The whole-genome sequence analyses have also identified loci under recent positive natural selection in the P. knowlesi genome, with evidence that different loci are affected in different populations. These provide a foundation to understand recent adaptation of the zoonotic parasite populations, and to track and interpret future changes as they emerge.